CN114926977B - Multitasking distribution method suitable for collaborative automatic driving - Google Patents
Multitasking distribution method suitable for collaborative automatic driving Download PDFInfo
- Publication number
- CN114926977B CN114926977B CN202210408057.XA CN202210408057A CN114926977B CN 114926977 B CN114926977 B CN 114926977B CN 202210408057 A CN202210408057 A CN 202210408057A CN 114926977 B CN114926977 B CN 114926977B
- Authority
- CN
- China
- Prior art keywords
- task
- vehicle
- function
- client
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 60
- 238000009826 distribution Methods 0.000 title abstract description 9
- 230000005540 biological transmission Effects 0.000 claims abstract description 59
- 230000006870 function Effects 0.000 claims abstract description 51
- 239000013598 vector Substances 0.000 claims abstract description 47
- 238000005265 energy consumption Methods 0.000 claims abstract description 34
- 238000004364 calculation method Methods 0.000 claims abstract description 23
- 239000011159 matrix material Substances 0.000 claims abstract description 22
- 238000005457 optimization Methods 0.000 claims abstract description 16
- 238000007726 management method Methods 0.000 claims abstract description 13
- 238000006243 chemical reaction Methods 0.000 claims abstract description 8
- 230000035945 sensitivity Effects 0.000 claims description 13
- 230000009466 transformation Effects 0.000 claims description 9
- 238000013459 approach Methods 0.000 claims description 7
- 238000004891 communication Methods 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 7
- 238000012886 linear function Methods 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 4
- 230000008878 coupling Effects 0.000 claims description 3
- 238000010168 coupling process Methods 0.000 claims description 3
- 238000005859 coupling reaction Methods 0.000 claims description 3
- 230000009467 reduction Effects 0.000 claims description 3
- 238000012545 processing Methods 0.000 description 6
- 238000004088 simulation Methods 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 2
- 230000007547 defect Effects 0.000 description 2
- 230000008447 perception Effects 0.000 description 2
- 238000007430 reference method Methods 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000001934 delay Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000670 limiting effect Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0108—Measuring and analyzing of parameters relative to traffic conditions based on the source of data
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L69/00—Network arrangements, protocols or services independent of the application payload and not provided for in the other groups of this subclass
- H04L69/16—Implementation or adaptation of Internet protocol [IP], of transmission control protocol [TCP] or of user datagram protocol [UDP]
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D30/00—Reducing energy consumption in communication networks
- Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Engineering & Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a multitask distribution method suitable for vehicle fog calculation in collaborative automatic driving, and belongs to the field of collaborative automatic driving. The invention combines vehicle fog calculation with the MIMO system and adopts analog beam forming. At the transmitter, the task data signals generated by each client vehicle are converted into parallel signal vectors through serial-parallel conversion, and then the transmission signal vectors are generated through a mixed precoding matrix. At the receiver, the transmitted signal vector is transformed and decoded by the channel matrix to obtain the final signal vector. And calculating the channel capacity of the antenna array, and modeling the multi-vehicle energy consumption and transmission rate optimization problem as a joint optimization problem function. The functions are subjected to constraint adding, and are converted into an approximate optimal solution model through convex set planning. And the management node calculates a task allocation method according to the model and issues and implements the task allocation method. The invention is suitable for the field of automatic driving, and can efficiently and parallelly unload tasks when the vehicle with multiple antennas cooperates with the automatic driving, thereby improving the data transmission rate and reducing the energy consumption.
Description
Technical Field
The invention relates to a multitask distribution method suitable for vehicle fog calculation in collaborative automatic driving, and belongs to the field of collaborative automatic driving.
Background
In recent years, in the field of collaborative autopilot of automobiles, companies in the world continuously increase the level of collaborative autopilot, gradually try to bring L2, L3, L4 class collaborative autopilot automobiles into the consumer market. For collaborative autopilot, if the algorithmic capability determines the upper limit of collaborative autopilot capability, then the vehicle gauge chip determines the lower limit. The lowest calculated power standard of the L5 level collaborative automatic driving is 500+TOPS, and the current great circle chip of the top has the great Orin x calculated power reaching 254TOPS, the domestic great circle chip has the calculated power of 72TOPS, and the supply is not required. So for most vehicles the calculation is far from satisfactory. For example, in 2019, the fourth quarter passenger car auto-park scenario, 92.8% of the vehicles select the ultrasonic scenario, and 7.2% of the vehicles select the ultrasonic + camera scenario. Although the latter is obviously better, the chip calculation force cannot support the image processing of the total of 288 frames/second of 8 cameras of the whole car, which has to be abandoned. Meanwhile, the energy consumption is greatly increased by cooperating with a high-performance chip and a redundant sensor required by automatic driving. Cloud computing is a main platform for analyzing and processing big data of the Internet of vehicles at present. However, cloud computing has many defects, and data transmission is cut far from a vehicle-mounted terminal, so that a large part of network bandwidth is occupied, network transmission delay is increased, emergency processing delay is caused, and a large amount of vehicles are overcrowded. in-Vehicle Fog Computing (VFC) concentrates communication and computing resources at the network edge, becoming a popular computing paradigm in-vehicle scenarios. The application of the vehicle fog calculation in the cooperative automatic driving can provide calculation force, reduce delay and energy consumption and effectively solve the problems. The key point of vehicle-mounted fog calculation is task unloading, and different task unloading technologies and methods determine energy consumption and delay of application implementation.
The collaborative automatic driving system comprises three parts, namely perception, decision planning and driving control, wherein perception data obtained by a vehicle sensor can be unloaded and sent to a service vehicle through a vehicle fog calculation task after being compressed by a processing algorithm, namely decision planning and driving control are carried out on a large-sized vehicle carrying a large amount of calculation resources. Collaborative autopilot functionality is evolving, from initial automatic emergency braking, to current AVP (Automated Valet Parking) automatic passenger parking, to future city piloting, which is creating more and more application tasks. In a collaborative autopilot scenario, the pulse proposition of application task processing is low latency and low energy consumption. Therefore, the parallel task allocation method has low time delay and low energy consumption and becomes a research focus. Alahmadi et al use a mixed integer linear programming model to determine the task allocation pattern to reduce the overall system energy consumption by reducing its average processing workload and network traffic, but this approach does not take into account service delays. Moreover, these theoretical studies and platform simulations merely stop the task offloading at the network level, without considering the configuration of the vehicle itself. As antenna manufacturing technology becomes more mature, antenna costs become lower and it is fully possible to install multiple wireless antennas on a single vehicle. The multiple antennas can effectively utilize channels, remarkably improve the data transmission rate and reduce the energy consumption. However, the transmission of different antennas on channels within the same frequency band can cause significant interference to each other. How to effectively distribute these different tasks in a vehicle fog computing network, and to enable more tasks to be transmitted in parallel while reducing communication interference is also a problem.
Disclosure of Invention
In order to overcome the defects of high energy consumption, high calculation force requirement and low data transmission rate in the prior art of collaborative automatic driving, the main aim of the invention is to provide a multi-task distribution method suitable for collaborative automatic driving, and by calculating vehicle fog, the efficient parallel task unloading of a vehicle with multiple antennae during collaborative automatic driving is realized, so that the data transmission rate is improved, and the energy consumption is reduced.
The invention aims at realizing the following technical scheme:
the invention discloses a multitask distribution method suitable for vehicle fog calculation in collaborative automatic driving, which combines the vehicle fog calculation with a MIMO system and adopts an analog beam forming technology. At the transmitter, the task data signals generated by each client vehicle are converted into parallel signal vectors through serial-parallel conversion, and then the transmission signal vectors are generated through a mixed precoding matrix. At the receiver, the transmitted signal vector is transformed and decoded by the channel matrix to obtain the final signal vector. And calculating the channel capacity of the antenna array, and modeling the multi-vehicle energy consumption and transmission rate optimization problem as a joint optimization problem function. The functions are subjected to constraint adding, and are converted into an approximate optimal solution model through convex set planning. In the application scene, the client vehicle sends a request, and the management node calculates a task allocation method according to the model and issues and implements the task allocation method. The method has the advantages that tasks are efficiently and parallelly unloaded when the vehicle with multiple antennae is in cooperation with automatic driving, so that the data transmission rate is improved, and the energy consumption is reduced.
The invention discloses a multitask distribution method suitable for collaborative automatic driving, which comprises the following steps:
step 1: modeling the non-convex non-concave optimization problem of the task allocation scheme as a solvable approximate optimization problem, and realizing effective reduction of energy consumption and delay by solving the approximate optimal task allocation scheme;
step 1.1, converting a plurality of task data streams generated by a single client vehicle into parallel signal vectors s through serial-parallel conversion i By the following constitutionI.e. the parallel signal vectors are passed through a digital precoding matrix and an analog precoding matrix to obtain a transmission vector x i I= { I } represents a client vehicle, j= { J } represents a server vehicle set, and K i = { k } represents the task generated by the client vehicle i, a ij = {0,1} represents the geographical association of the client vehicle i with the service vehicle j, the value of j being 1, +.>Indicating whether task k is assigned to service vehicle j, if the value is 1, it is 0 and it is not assigned, G i Digital precoding matrix for vehicle i, B i For the analog precoding matrix of vehicle i, s i ＝[s1,s2,···,s|K i |]Representing parallel signal vectors generated by vehicle i, s k Signals representing task k, g k Digital precoding vector, x, representing task k i A transmission signal vector representing a vehicle i;
step 1.2: transmission vector x i The signal vector received by the service vehicle can be obtained through channel matrix transformation Decoding to obtain signal vector-> Consists of three parts, namely an expected signal of a service vehicle, other interference signals and noise interference. According to->The antenna array channel capacity for task k between client vehicle i and service vehicle j can be calculated asWherein sigma 2 Is the variance of Gaussian white noise, is a constant; the K 'with' refers to the set K 'of K after removing the ith task' i Is a task in (a). Setting up to use a uniform digital precoding method, i.e. g, for each generated task k Let h =g ij ＝|f j ·H ij ·g·B i 2, then simplify->Is that Representing the transmission power allocated to task k, f j Is the analog beamforming vector, H, used to serve vehicle j ij Representing the channel matrix between client i and server j +.>Representing customersAntenna array channel capacity for task k between vehicle i and service vehicle j;
step 1.3: setting a sensitivity variable alpha representing the data rate, the problem function P1 is optimized as
Wherein alpha is E [0,1 ]]Weights indicating the degree of sensitivity to rate, it being apparent that the greater alphaThe greater the influence of (a) is on the simultaneous power p k The smaller the impact of (2); 1a represents the task k power consumption range, +.>Representing minimum and maximum power consumed by the transmission task k; 1b indicates that the service vehicle j can at most serve eta j Task eta j Load capacity for service vehicle j; 1c represents a transmission data rate constraint, where τ k Is the minimum transmission data rate for task k.
Step 1.4: to ensure that the problem function and constraint function are convex functions, one must solveAnd p k The problem of coupling these 2 variables, for which a new variable +.> And introducing 2 new constraints (2 a) and (2 b) to ensure +.>In any case can be equal to +.>And the function does not contain the part of the multiplication of the two,/->All possible values are as follows:
additional constraints:
by this transformation, the P1 equivalence transformation becomes P2:
wherein the method comprises the steps ofThe method comprises the following steps:
step 1.5: by x k Instead ofIs defined by the variables i and j, k being the basis of the quantity, h being denoted by h ij Wherein i is
And j is a variable. Then solve for P2 using convex set (d.c.) programming, P2 translates into P3:
P3:
wherein:
step 1.6: the data transmission rate is approximated by a linear function in a differential manner, and there are:
the equation is converted into a log- (kx+b) form, so that the unique concave-convex performance of the function is ensured, an approximate optimal solution can be obtained through iteration, and then P3 is converted into P4:
where l is the number of iterations, because the solution of the convex difference always approaches the optimal solution, the optimal solution point that is approximated by the linear function iteration approaches the optimal solution point of the original function. Wherein the method comprises the steps ofThe value for the last iteration is a determined number. Initially feasible solution->By using Q (x) k ) And (3) solving a convex optimization problem for the objective function. And when the l reaches a preset maximum value, the iteration is terminated.
Step 2: aiming at different calculation demands of different data, the vehicle generates a multi-task data stream, and starts to send task unloading demand information to a management node in a communication range, wherein the task unloading demand information comprises delay demands, data quantity and data types;
step 3: the vehicle operation region is divided into hexagonal grids, and a base station closest to the center in each grid is used as a management node, and the management node obtains the information of service vehicles in a range through broadcasting;
step 4: the management node refers to the model established in the step 1, and task unloading decision is carried out by integrating task information and service vehicle information, and the specific steps are as follows:
step 4 a): taking a client vehicle set, a service vehicle set and an unassigned task set as input;
step 4 b): setting a variable alpha according to the sensitivity of the application to the data rate, setting iteration index to 0, maximum iteration times, convergence tolerance and other parameters;
step 4 c): the function P3 referenced in step 1 to maximize the overall task data transfer rate and the energy consumption average can be expressed as a solveable function Q (x k ) Minus a non-convex non-concave function Z (x k ) By applying a non-convex non-concave function Z (x k ) Setting zero to obtain an initial allocation decision set x0, wherein x0 is taken as a standard allocation decision set;
step 4 d): by applying a function Z (x k ) Linearizing to obtain Z (x) k ) The function of maximizing the total task data transmission rate and the energy consumption mean value is expressed as a solvable convex function P4 with respect to the overall upper estimator of the first-order Taylor expansion of the iteration times; step 4 e): bringing the standard allocation decision set into P4 to obtain a new allocation decision set, taking the new allocation decision set as the standard allocation decision set, adding 1 to the iteration index, and recalculating according to the step 4 e) if the convergence degree is greater than the preset convergence tolerance or the iteration index is less than the preset maximum iteration number;
step 5: according to the step 4, an optimal task allocation decision set is obtained, and the set is sent to a client vehicle;
step 6: the client vehicle performs serial-parallel conversion on the task data stream into parallel signal vectors;
step 7: digital precoding and analog precoding are carried out on the parallel signal vectors, and transmission signal vectors are generated and transmitted through multiple antennas;
step 8: the transmission signal vector is transmitted to the service vehicle through a beam forming technology, and the service vehicle starts to calculate the task after decoding and receiving the designated task data.
Step 9: the calculated decision scheme is transmitted back to the client vehicle, and the client vehicle starts decision execution, so that the task is efficiently and parallelly unloaded when the vehicle with multiple antennas is in cooperative automatic driving, the data transmission rate is improved, and the energy consumption is reduced.
The beneficial effects are that:
1. compared with other task unloading methods, the multi-task allocation method suitable for collaborative automatic driving disclosed by the invention has the advantages that the MIMO system is utilized to improve the channel capacity, the multi-antenna array is applied to vehicle-mounted fog calculation, the concurrent calculation problem is solved by adopting a convex optimization method, and meanwhile, a plurality of task flows are received and transmitted, so that the data transmission rate is effectively improved, and the vehicle can have a faster response speed.
2. Compared with the similar task unloading method using multiple antennas, the multi-task allocation method suitable for collaborative automatic driving disclosed by the invention realizes a more reasonable allocation scheme through a mixed precoding and convex optimization method, so that the data transmission rate is faster, the energy consumption is greatly reduced, and the vehicle has longer endurance.
Drawings
FIG. 1 is a flow chart of a method of multitasking for coordinated autopilot in accordance with the present disclosure;
fig. 2 is a transmission rate comparison graph of a multitasking method and a reference task allocation method for different rate sensitivities α in a collaborative autopilot scenario.
Fig. 3 is an energy consumption comparison graph of a multi-tasking method and a reference tasking method for different speed sensitivities α in a collaborative autopilot scenario.
Detailed Description
The present invention will be described in detail with reference to the accompanying drawings and examples. The technical problems and the beneficial effects solved by the technical proposal of the invention are also described, and the described embodiment is only used for facilitating the understanding of the invention and does not have any limiting effect.
First, a MIMO system will be described. The MIMO system (Multiple-Input Multiple-Output) refers to a system in which a plurality of transmitting antennas and receiving antennas are used at a transmitting end and a receiving end, respectively, so that signals are transmitted and received through the plurality of antennas at the transmitting end and the receiving end, thereby improving communication quality. The system can fully utilize space resources, realize multiple transmission and multiple reception through a plurality of antennas, and can doubly improve the system channel capacity under the condition of not increasing frequency spectrum resources and antenna transmitting power.
In the application scenario of collaborative automatic driving, in order to improve the frequency utilization rate and increase the capacity of the system, a frequency multiplexing technology is often adopted, and this can bring about the problem of co-channel interference. Co-channel interference, that is, interference caused by a receiver receiving a co-channel useful signal, means that the carrier frequency of the useful signal is the same as the carrier frequency of the useful signal. The invention aims to solve the problems that a large amount of calculation demands are often generated by collaborative automatic driving applications, such as APA automatic parking assistance, AVM panoramic looking around monitoring, TJP traffic jam navigation and the like, a large amount of data are required to be transmitted and processed, and meanwhile, the data transmission rate of collaborative automatic driving application task unloading is improved by using a MIMO system to enable vehicle response to be more sensitive, reduce energy consumption and increase vehicle endurance, and simultaneously, the same-frequency interference problem and noise problem brought by the MIMO system are also required to be reduced because of the sensitivity of application scenes to time.
The embodiment is a specific application of the multi-task allocation method suitable for collaborative automatic driving in an automatic driving scene. In this embodiment, the actual track of the bus in the square kilometer area of helsinki 1 during the period 08:00-08:05 of 2018, 9 and 18 is collected, the latitude of the area is 60 degrees 1201500-60 degrees 1205000, the longitude is 24 degrees 53041500-24 degrees 5402400, and 240 vehicle tracks in that period are generated by using SUMO simulation software, wherein 50% of vehicles are considered to have cooperative automatic driving calculation requirements, and task unloading is needed. The task unloading method comprises the following specific steps:
step 1: modeling the non-convex non-concave optimization problem of the task allocation scheme as a solvable approximate optimization problem, and realizing effective reduction of energy consumption and delay by solving the approximate optimal task allocation scheme;
step 1.1: multiple task data streams generated by single client vehicles are converted into parallel signal vectors s through serial-parallel conversion i By the following constitutionI.e. the parallel signal vectors are passed through a digital precoding matrix and an analog precoding matrix to obtain a transmission vector x i I= { I } represents a client vehicle, j= { J } represents a server vehicle set, and K i = { k } represents the task generated by the client vehicle i, a ij = {0,1} represents the geographical association of the client vehicle i with the service vehicle j, the value of j being 1, +.>Indicating whether task k is assigned to service vehicle j, if the value is 1, it is 0 and it is not assigned, G i Digital precoding matrix for vehicle i, B i For the analog precoding matrix of vehicle i, s i ＝[s1,s2,···,s|K i |]Representing parallel signal vectors generated by vehicle i, s k Signals representing task k, g k Digital precoding vector, x, representing task k i A transmission signal vector representing a vehicle i;
step 1.2: transmission vector x i The signal vector received by the service vehicle can be obtained through channel matrix transformation Decoding to obtain signal vector-> Consists of three parts, namely an expected signal of a service vehicle, other interference signals and noise interference. According to->The antenna array channel capacity for task k between client vehicle i and service vehicle j can be calculated asWherein sigma 2 Is the variance of Gaussian white noise, is a constant; the K 'with' refers to the set K 'of K after removing the ith task' i Is a task in (a). Setting up to use a uniform digital precoding method, i.e. g, for each generated task k Let h =g ij ＝|f j ·H ij ·g·B i 2, then simplify->Is that Representing the transmission power allocated to task k, f j Is the analog beamforming vector, H, used to serve vehicle j ij Representing the channel matrix between client i and server j +.>Antenna array channel capacity representing task k between client vehicle i and service vehicle j;
step 1.3: setting a sensitivity variable alpha representing the data rate, the problem function P1 is optimized as
Wherein alpha is E [0,1 ]]Weights indicating the degree of sensitivity to rate, it being apparent that the greater alphaThe greater the influence of (a) is on the simultaneous power p k The smaller the impact of (2); 1a represents the task k power consumption range, +.>Representing minimum and maximum power consumed by the transmission task k; 1b indicates that the service vehicle j can at most serve eta j Task eta j Load capacity for service vehicle j; 1c represents a transmission data rate constraint, where τ k Is the minimum transmission data rate for task k.
Step 1.4: to ensure that the problem function and constraint function are convex functions, one must solveAnd p k The problem of coupling these 2 variables, for which a new variable +.> And introducing 2 new constraints (2 a) and (2 b) to ensure +.>In any case can be equal to +.>And the function does not contain the part of the multiplication of the two,/->All possible values are as follows:
additional constraints:
by this transformation, the P1 equivalence transformation becomes P2:
wherein the method comprises the steps ofThe method comprises the following steps:
step 1.5: by x k Instead ofIs defined by the variables i and j, k being the basis of the quantity, h being denoted by h ij Wherein i is
And j is a variable. Then solve for P2 using convex set (d.c.) programming, P2 translates into P3:
P3:
wherein:
step 1.6: the data transmission rate is approximated by a linear function in a differential manner, and there are:
the equation is converted into a log- (kx+b) form, so that the unique concave-convex performance of the function is ensured, an approximate optimal solution can be obtained through iteration, and then P3 is converted into P4:
where l is the number of iterations, because the solution of the convex difference always approaches the optimal solution, the optimal solution point that is approximated by the linear function iteration approaches the optimal solution point of the original function. Wherein the method comprises the steps ofThe value for the last iteration is a determined number. Initially feasible solution->By using Q (x) k ) And (3) solving a convex optimization problem for the objective function. And when the l reaches a preset maximum value, the iteration is terminated.
Step 2: after calculation, 15 buses and 120 automobiles appearing in the simulation become service vehicles and customer vehicles respectively, and for the calculation requirements of the application, it is assumed that each customer vehicle will generate 3 tasks. Starting to send task unloading requirement information to management nodes in a communication range;
the driver selects a certain cooperative automatic driving function, such as urban navigation, the client vehicle starts to collect data through a camera, an ultrasonic radar, a millimeter wave radar and other sensors, vehicle state information is generated in real time, and meanwhile, the driver may start an entertainment function to generate related data.
Step 3: in order to secure the transmission data rate, the effective communication distance of the service vehicle is set to 100 meters. In addition, the management node obtains information of the service vehicle in the range by broadcasting, including position and load information, and in the simulation performed, the load capacity eta of the service vehicle j The values of (2) vary from 18 to 69 to ensure that the iteration problem is solvable;
step 4: the management node refers to the model established in the step 1, and task unloading decision is carried out by integrating task information and service vehicle information, and the specific steps are as follows:
step 4 a): taking a client vehicle set i=120, a service vehicle set j=5 and an unassigned task set as input;
step 4 b): setting the power range consumed by the transmit task to [10dBm,23dBm](i.e., [0.01W,0.2W ]]) And uses 0.1W as the initial transmit power. The antenna array channel parameter values of all the associated pairs of the client vehicle and the service vehicle can be obtained through multiple iterations, and h ij =2046. In the simulation, assuming that the antenna array of each vehicle is omni-directional, the variable σ2=0.1, the iteration index is set to 0, and the maximum number of iterations is set to 30. Scalar weight a refers to the trade-off between transmission data rate and power consumption. The smaller α, the more power sensitive the proposed strategy is and vice versa. Adjusting α from 0.2 to 1 in steps of 0.2;
step 4 c): the function P3 referenced in step 1 to maximize the overall task data transfer rate and the energy consumption average can be expressed as a solveable function Q (x k ) Minus a non-convex non-concave function Z (x k ) By applying a non-convex non-concave function Z (x k ) Setting zero to obtain an initial allocation decision set x0, wherein x0 is taken as a standard allocation decision set;
step 4 d): by applying a function Z (x k ) Linearizing to obtain Z (x) k ) The function of maximizing the total task data transmission rate and the energy consumption mean value is expressed as a solvable convex function P4 with respect to the overall upper estimator of the first-order Taylor expansion of the iteration times; step 4 e): bringing the standard allocation decision set into P4 to obtain a new allocation decision set, taking the new allocation decision set as the standard allocation decision set, adding 1 to the iteration index, and recalculating according to the step 4 e) if the convergence degree is greater than the preset convergence tolerance or the iteration index is less than the preset maximum iteration number;
step 5: according to the step 4, an optimal task allocation decision set is obtained, and the set is sent to a client vehicle;
step 6: the client vehicle performs serial-parallel conversion on the task data stream into parallel signal vectors;
step 7: digital precoding and analog precoding are carried out on the parallel signal vectors, and transmission signal vectors are generated and transmitted through multiple antennas;
step 8: the transmission signal vector is transmitted to the service vehicle through a beam forming technology, and the service vehicle starts to calculate the task after decoding and receiving the designated task data.
Step 9: and transmitting the calculated decision scheme back to the client vehicle, and starting decision execution by the client vehicle.
For comparison, simulations implement a well-established task allocation method being applied, called the benchmark task allocation method, in which a probability distribution method is used to balance task allocation operations with load operations running on the service vehicle. In the reference task allocation method, the more tasks currently carried by the service vehicle, the lower the probability of new tasks being allocated to the service vehicle. Specifically, the probability of assigning a task to a service vehicle is:
fig. 2 is a transmission rate comparison graph of a task allocation method and a reference task allocation method of different rate sensitivities α in a cooperative automatic driving scenario. The abscissa is the reference task allocation method, the α is 0.4,0.6 and the inventive method of 0.8, and the ordinate is the comparison of the transmission power with the reference task allocation method. The average transmission data rate increases with increasing α. When alpha=1, the maximum data rate can reach 2.07bps/Hz, and when alpha is 0.4,0.6,0.8, the transmission data rate is 5.8%, 6.4% and 8.2% higher than the reference method respectively.
Fig. 3 is an energy consumption comparison diagram of a task allocation method and a reference task allocation method of different speed sensitivities α in a cooperative automatic driving scenario. The abscissa is the reference task allocation method, the α is 0.4,0.6 and the method of the invention is 0.8, and the ordinate is the comparison of the energy consumption with the reference task allocation method. The average energy consumption increases with increasing alpha. When α=0.2, the minimum power consumption can reach 0.03W. As alpha changes from 0.8 to 1, the average energy consumption increases significantly. This is because when α=1, the average energy consumption can almost reach its upper limit. When alpha is 0.4 and alpha is 0.8, the energy consumption is reduced by 38.3 percent and 23.3 percent respectively compared with the reference method.
According to fig. 2, fig. 3 shows that the highest transmission data rate is obtained when α is 0.8, and the lowest power consumption is obtained when α is 0.4. Therefore, when the application task is particularly sensitive to delay, we recommend that alpha be set to 0.8, and when the application task is not sensitive to delay, we recommend that alpha be set to 0.4, so that energy can be effectively saved.
The method is suitable for a convex optimization-based multi-task distribution method for collaborative automatic driving vehicle fog calculation, effectively utilizes vehicle-mounted multi-antennae in a vehicle fog calculation network to carry out parallel task unloading, and can furthest improve the data transmission rate and reduce the energy consumption.
The foregoing detailed description has set forth the objects, aspects and advantages of the invention in further detail, it should be understood that the foregoing description is only illustrative of the invention and is not intended to limit the scope of the invention, but is to be accorded the full scope of the invention as defined by the appended claims.
Claims (2)
1. A multitasking method suitable for collaborative autopilot, characterized by: comprises the following steps of the method,
step 1: modeling the non-convex non-concave optimization problem of the task allocation scheme as a solvable approximate optimization problem, and realizing effective reduction of energy consumption and delay by solving the approximate optimal task allocation scheme;
the implementation method of the step 1 is that,
step 1.1, converting a plurality of task data streams generated by a single client vehicle into parallel signal vectors s through serial-parallel conversion i By the following constitutionI.e. the parallel signal vectors are passed through a digital precoding matrix and an analog precoding matrix to obtain a transmission vector x i I= { I } represents a client vehicle, j= { J } represents a server vehicle set, and K i = { k } represents the task generated by the client vehicle i, a ij = {0,1} represents the geographical association of the client vehicle i with the service vehicle j, the value of j being 1, +.>Indicating whether task k is assigned to service vehicle j, if the value is 1, it is 0 and it is not assigned, G i Digital precoding matrix for vehicle i, B i For the analog precoding matrix of vehicle i, s i ＝[s1,s2,···,s|K i |]Representing parallel signal vectors generated by vehicle i, s k Signals representing task k, g k Digital precoding vector, x, representing task k i A transmission signal vector representing a vehicle i;
step 1.2: transmission signal vector x i The signal vector received by the service vehicle can be obtained through channel matrix transformation Decoding to obtain signal vector-> The system consists of an expected signal, other interference signals and noise interference of a service vehicle; according to->The antenna array channel capacity for task k between client vehicle i and service vehicle j can be calculated asWherein sigma 2 Is the variance of Gaussian white noise, is a constant; the K 'with' refers to the set K 'of K after removing the ith task' i Is a task in (1); setting up to use a uniform digital precoding method, i.e. g, for each generated task k Let h =g ij ＝|f j ·H ij ·g·B i | 2 Then simplify->Is that Representing the transmission power allocated to task k, f j Is the analog beamforming vector, H, used to serve vehicle j ij Representing the channel matrix between client i and server j +.>Antenna array channel capacity representing task k between client vehicle i and service vehicle j;
step 1.3: setting a sensitivity variable alpha representing the data rate, the problem function P1 is optimized as
s.t.
Wherein alpha is E [0,1 ]]Weights indicating the degree of sensitivity to rate, it being apparent that the greater alphaThe greater the influence of (a) is on the simultaneous power p k The smaller the impact of (2); 1a represents the task k power consumption range, +.>Representing the minimum and maximum work consumed by a transfer task kA rate; 1b indicates that the service vehicle j can at most serve eta j Task, eta j Load capacity for service vehicle j; 1c represents a transmission data rate constraint, where τ k Minimum transmission data rate for task k;
step 1.4: to ensure that the problem function and constraint function are convex functions, one must solveAnd p k The problem of coupling these 2 variables, for which a new variable +.>And introducing 2 new constraints (2 a) and (2 b) to ensure +.>In any case can be equal to +.>And the function does not contain the part of the multiplication of the two,/->All possible values are as follows:
additional constraints:
by this transformation, the P1 equivalence transformation becomes P2:
wherein the method comprises the steps ofThe method comprises the following steps:
step 1.5: by x k Instead ofIs defined by the variables i and j, k being the basis of the quantity, h being denoted by h ij Wherein i and j are variables; then solve for P2 using convex set (d.c.) programming, P2 translates into P3:
P3:
wherein:
step 1.6: the data transmission rate is approximated by a linear function in a differential manner, and there are:
the equation is converted into a log- (kx+b) form, so that the unique concave-convex performance of the function is ensured, an approximate optimal solution can be obtained through iteration, and then P3 is converted into P4:
wherein l is the iteration number, and because the solution of the convex difference can always approach the optimal solution, the optimal solution point which is approximated by the iteration of the linear function approaches the optimal solution point of the original function; wherein the method comprises the steps ofThe value for the last iteration is a determined number; initially feasible solution->By using Q (x) k ) The convex optimization problem for solving the objective function is obtained; when l reaches a preset maximum value, the iteration is terminated;
step 2: aiming at different calculation demands of different data, the vehicle generates a multi-task data stream, and starts to send task unloading demand information to a management node in a communication range, wherein the task unloading demand information comprises delay demands, data quantity and data types;
step 3: the vehicle operation region is divided into hexagonal grids, and a base station closest to the center in each grid is used as a management node, and the management node obtains the information of service vehicles in a range through broadcasting;
step 4: the management node refers to the model established in the step 1, and task unloading decision is carried out through comprehensive task information and service vehicle information;
step 5: according to the step 4, an optimal task allocation decision set is obtained, and the set is sent to a client vehicle;
step 6: the client vehicle performs serial-parallel conversion on the task data stream into parallel signal vectors;
step 7: digital precoding and analog precoding are carried out on the parallel signal vectors, and transmission signal vectors are generated and transmitted through multiple antennas;
step 8: the transmission signal vector is transmitted to the service vehicle through a beam forming technology, and the service vehicle starts to calculate the task after decoding and receiving the designated task data;
step 9: the calculated decision scheme is transmitted back to the client vehicle, and the client vehicle starts decision execution, so that the task is efficiently and parallelly unloaded when the vehicle with multiple antennas is in cooperative automatic driving, the data transmission rate is improved, and the energy consumption is reduced.
2. A method of multitasking for a cooperative autopilot in accordance with claim 1 wherein: the implementation method of the step 4 is that,
step 4 a): taking a client vehicle set, a service vehicle set and an unassigned task set as input;
step 4 b): setting a variable alpha according to the sensitivity of the application to the data rate, setting iteration index to 0, maximum iteration times, convergence tolerance and other parameters;
step 4 c): the function P3 referenced in step 1 to maximize the overall task data transfer rate and the energy consumption average can be expressed as a solveable function Q (x k ) Minus a non-convex non-concave function Z (x k ) By applying a non-convex non-concave function Z (x k ) Setting zero to obtain an initial allocation decision set x0, wherein x0 is taken as a standard allocation decision set;
step 4 d): by applying a function Z (x k ) Linearizing to obtain Z (x) k ) The function of maximizing the total task data transmission rate and the energy consumption mean value is expressed as a solvable convex function P4 with respect to the overall upper estimator of the first-order Taylor expansion of the iteration times; step 4 e): bringing the standard allocation decision set into P4 to obtain a new allocation decision set, taking the new allocation decision set as the standard allocation decision set, adding 1 to the iteration index, and recalculating according to the step 4 e) if the convergence degree is larger than the preset convergence tolerance or the iteration index is smaller than the preset maximum iteration number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210408057.XA CN114926977B (en) | 2022-04-19 | 2022-04-19 | Multitasking distribution method suitable for collaborative automatic driving |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210408057.XA CN114926977B (en) | 2022-04-19 | 2022-04-19 | Multitasking distribution method suitable for collaborative automatic driving |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114926977A CN114926977A (en) | 2022-08-19 |
| CN114926977B true CN114926977B (en) | 2024-04-05 |
Family
ID=82807131
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210408057.XA Active CN114926977B (en) | 2022-04-19 | 2022-04-19 | Multitasking distribution method suitable for collaborative automatic driving |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114926977B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104821838A (en) * | 2015-04-24 | 2015-08-05 | 浙江理工大学 | Energy efficiency maximization-based multi-user information and energy simultaneous transmission transceiver design method |
| US9883511B1 (en) * | 2012-12-05 | 2018-01-30 | Origin Wireless, Inc. | Waveform design for time-reversal systems |
| CN111614419A (en) * | 2020-01-10 | 2020-09-01 | 南京邮电大学 | NOMA-based high-safety unloading resource allocation method for mobile edge computing network task |
| CN112567673A (en) * | 2018-08-09 | 2021-03-26 | 康维达无线有限责任公司 | Beamforming and grouping for NR V2X |
| CN113721198A (en) * | 2021-09-09 | 2021-11-30 | 哈尔滨工程大学 | Physical layer security combined beam forming method for dual-function MIMO radar communication system |
| CN113743469A (en) * | 2021-08-04 | 2021-12-03 | 北京理工大学 | Automatic driving decision-making method fusing multi-source data and comprehensive multi-dimensional indexes |
| CN113904947A (en) * | 2021-11-15 | 2022-01-07 | 湖南大学无锡智能控制研究院 | Vehicle-road cooperative distributed edge computing task unloading and resource allocation method and system |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10805022B2 (en) * | 2018-01-12 | 2020-10-13 | The Euclide 2012 Investment Trust | Method of using time domain subspace signals and spatial domain subspace signals for location approximation through orthogonal frequency-division multiplexing |
-
2022
- 2022-04-19 CN CN202210408057.XA patent/CN114926977B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9883511B1 (en) * | 2012-12-05 | 2018-01-30 | Origin Wireless, Inc. | Waveform design for time-reversal systems |
| CN104821838A (en) * | 2015-04-24 | 2015-08-05 | 浙江理工大学 | Energy efficiency maximization-based multi-user information and energy simultaneous transmission transceiver design method |
| CN112567673A (en) * | 2018-08-09 | 2021-03-26 | 康维达无线有限责任公司 | Beamforming and grouping for NR V2X |
| CN111614419A (en) * | 2020-01-10 | 2020-09-01 | 南京邮电大学 | NOMA-based high-safety unloading resource allocation method for mobile edge computing network task |
| CN113743469A (en) * | 2021-08-04 | 2021-12-03 | 北京理工大学 | Automatic driving decision-making method fusing multi-source data and comprehensive multi-dimensional indexes |
| CN113721198A (en) * | 2021-09-09 | 2021-11-30 | 哈尔滨工程大学 | Physical layer security combined beam forming method for dual-function MIMO radar communication system |
| CN113904947A (en) * | 2021-11-15 | 2022-01-07 | 湖南大学无锡智能控制研究院 | Vehicle-road cooperative distributed edge computing task unloading and resource allocation method and system |
Non-Patent Citations (1)
| Title |
|---|
| 基于数模混合架构的波束赋形技术研究;张砚秋;《中国优秀硕士学位论文全文数据库信息科技辑》(第01期);I136-1039 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114926977A (en) | 2022-08-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Cheng et al. | Air-ground integrated mobile edge networks: Architecture, challenges, and opportunities | |
| He et al. | 6G cellular networks and connected autonomous vehicles | |
| CN113904947B (en) | Vehicle-road cooperative distributed edge computing task unloading and resource allocation method and system | |
| CN110650457B (en) | Joint optimization method for task unloading calculation cost and time delay in Internet of vehicles | |
| CN111464976A (en) | Vehicle task unloading decision and overall resource allocation method based on fleet | |
| CN110928658A (en) | Cooperative task migration system and algorithm of vehicle-side cloud cooperative architecture | |
| CN113163365B (en) | Unmanned aerial vehicle support networking resource optimization method based on alternating direction multiplier algorithm | |
| CN109120552A (en) | Bandwidth and power multiple target cross-layer optimizing method towards QOS in a kind of AOS | |
| CN112272232B (en) | Millimeter wave Internet of vehicles resource scheduling method and device, electronic equipment and storage medium | |
| CN114980169A (en) | Unmanned aerial vehicle auxiliary ground communication method based on combined optimization of track and phase | |
| CN117498900A (en) | Resource allocation device and method for honeycomb-removing large-scale MIMO (multiple input multiple output) general sense integrated system | |
| Yu et al. | Edge-assisted collaborative perception in autonomous driving: A reflection on communication design | |
| CN116866933A (en) | Unmanned aerial vehicle assisted edge computing network resource allocation method based on intelligent reflection surface assistance | |
| CN114926977B (en) | Multitasking distribution method suitable for collaborative automatic driving | |
| CN114153515A (en) | Expressway internet of vehicles task unloading algorithm based on 5G millimeter wave communication | |
| CN116709249A (en) | Management method for edge calculation in Internet of vehicles | |
| Xiong et al. | Joint connection modes, uplink paths and computational tasks assignment for unmanned mining vehicles’ energy saving in mobile edge computing networks | |
| CN111132298B (en) | Power distribution method and device | |
| CN115964178A (en) | Internet of vehicles user computing task scheduling method and device and edge service network | |
| CN115665805A (en) | Point cloud analysis task-oriented edge computing resource scheduling method and system | |
| CN114885351A (en) | Heterogeneous network multi-target optimization interference method and device and terminal equipment | |
| CN110831204A (en) | Power distribution method and system for downlink of high-speed moving train | |
| Wang et al. | Latest advances in spectrum management for 6g communications | |
| CN114666766B (en) | Internet of things gateway communication load sharing method and system | |
| CN114363855A (en) | Efficient power distribution method in Internet of vehicles |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |